Path: utzoo!attcan!uunet!super!rminnich From: rminnich@super.ORG (Ronald G Minnich) Newsgroups: comp.protocols.tcp-ip Subject: Re: YP, /etc/hosts, domains Message-ID: <5004@super.ORG> Date: 17 Jan 89 17:50:23 GMT References: Sender: news@super.ORG Reply-To: rminnich@super.UUCP (Ronald G Minnich) Organization: Supercomputing Research Center, Lanham, Md. Lines: 32 In article mrc@SUMEX-AIM.STANFORD.EDU (Mark Crispin) writes: > In other words, I'd dearly love to turn YP off, if only I could get >ordinary domain lookups working. How do I do this? Do I have to run a domain >server too, just to get the resolver working? well, i had this problem too. I have two possible solutions: 1) You got yp source? Then build a hot-wired ypbind/ypserv pair that only look at your machines IP address. Bingo, you have your own private servers, but the domain name is still the same as your networks (important for lots of things). This works well here. 2) No yp source? Set up your own yp domain, like crispin.org or something, but this is messier; you have to set domainname and all that and make sendmail work right. I don't like this approach too much. 3) Fix up /etc/rc so that it does a ypset to your host, and have your own maps built for only your host, and get rid of all but the host.* maps. This gets messy too, and you have to keep making sure that (1) only you and no one else binds to your host (2) you STAY bound to your host (ypbind will move you around as it sees fit) after messing with (1), (2), and (3), i stuck with (1). Note that there are several funny bugs in some ypservs (Ultrix comes to mind) so you have to check out the source of your source. Specific example: yp uses dbm, not ndbm, but ypserv calls dbm_close when it switches maps (with a char * yet!!)(should be DBM *;found in xinu & ultrix) Newest sun source shows it calling dbmclose, but still with a char * argument. (dbmclose takes no arguments) If you want c-diffs for what i had to do here let me know. I am pretty sure Stanford has a source license, right :-)? ron rminnich@super.org P.S. yp to me always seems to have been a good first cut, but i think it is time for another go at the whole problem. hesiod, anyone?